A Dimeric Titanium-Containing Polyoxometalate. Synthesis

31P NMR reveals that H1 (δ −12.76) readily forms in MeCN from the Keggin ... The addition of H2O2 to H1 or H2 results in the formation of a peroxo ...
0 downloads 0 Views 155KB Size
Download PDF
3828

Inorg. Chem. 2000, 39, 3828-3837

A Dimeric Titanium-Containing Polyoxometalate. Synthesis, Characterization, and Catalysis of H2O2-Based Thioether Oxidation Oxana A. Kholdeeva,*,† Gennadii M. Maksimov,† Raisa I. Maksimovskaya,† Lyubov A. Kovaleva,† Martin A. Fedotov,† Vladimir A. Grigoriev,‡ and Craig L. Hill‡,* Department of Chemistry, Emory University, 1515 Pierce Drive, Atlanta, Georgia 30322, and †Boreskov Institute of Catalysis, Prospekt Ak. Lavrentieva 5, Novosibirsk 630090, Russia ReceiVed January 20, 2000

The previously unknown titanium(IV)-containing µ-hydroxo dimeric heteropolytungstate (Bu4N)7[(PTiW11O39)2OH] (TBA salt of H1) has been synthesized, starting from H5PTiW11O40, and characterized by elemental analysis, multinuclear (31P, 17O, 183W) NMR, IR, FAB-MS, cyclic voltammetry, and potentiometric titration. 31P NMR reveals that H1 (δ -12.76) readily forms in MeCN from the Keggin monomer (POM), PTiW11O405- (2, δ -13.34), upon the addition of 1.5 equiv of H+, via the protonated species, P(TiOH)W11O394- (H2, δ -13.44). The ratio of H1, 2, and H2, which are present in equilibrium in MeCN solution at 25 °C, depends on the concentration of both H+ and H2O. The Ti-O-Ti linkage readily reacts with nucleophilic reagents, such as H2O and ROH, to yield monomeric Keggin derivatives. µ-Hydroxo dimer H1 shows higher catalytic activity than 2 for thioether oxidation by hydrogen peroxide in acetonitrile. The reaction proceeds readily at room temperature and affords the corresponding sulfoxide and sulfone in ca. quantitative yield. The addition of H2O2 to H1 or H2 results in the formation of a peroxo complex, most likely the hydroperoxo complex P(TiOOH)W11O394- (I), which has 31P NMR resonance at -12.43 ppm. The rate of the formation of I is higher from H2 than from H1. When H1 is used as a catalyst precursor, the rates of the thioether oxidation and peroxo complex formation increase with increasing H2O concentration, which favors the cleavage of H1 to H2. H2O2 in MeCN slowly converts 2 to another peroxotitanium complex, P(TiO2)W11O395- (II), which has 31P NMR resonance at -12.98 ppm. Peroxo complexes I and II differ in their protonation state and interconvert fast on the 31P NMR time scale. Addition of 1 equiv of H+ completely converts II to I, while 1 equiv of OH- completely converts I to II. 31P NMR confirms that I is stable under turnover conditions (thioether, H2O2, MeCN). Contrary to two-phase systems such as dichloroethane/ aqueous H2O2, no products resulting from the destruction of the Keggin POM were detected in MeCN in the presence of H2O2 (a 500-fold molar excess). The reactivity of I, generated in situ from II by adding 1 equiv of H+, toward organic sulfides under stoichiometric conditions was confirmed using both 31P NMR and UV-vis spectroscopy. This is a rare demonstration of the direct stoichiometric oxidation of an organic substrate by a titanium peroxo complex.

Introduction The selective catalytic oxidation of organic compounds is of both academic and industrial interest.1-5 d-Electron-transitionmetal-substituted polyoxometalates (POMs), so-called inorganic metalloporphyrins, have attracted much attention as oxidation catalysts because they are thermodynamically stable against oxidative degradation, unlike metalloporphyrins and other metal complexes with organic ligands; yet their d-metal “active sites” can be extensively modified synthetically.6-10 These compounds

combine the selectivity advantages of homogeneous catalysts with the stability advantages of heterogeneous catalysts. Since they consist of a closely packed array of oxide anions, they can be regarded as discrete fragments of extended metal oxide lattices and, thus, can function as tractable homogeneous probes of heterogeneous catalytic oxidation reaction mechanisms.6,7,10,11 Much work has been reported on these complexes as catalysts for the oxidation of organic compounds by various oxygen donors, such as PhIO,12-16 H2O2,8,17-24 t-BuOOH,15,25 NaIO4,26

* Corresponding authors. O.A.K.: fax, (3832) 343766; e-mail, khold@ catalysis.nsk.su. C.L.H.: fax, 404-727-6076; e-mail, [email protected]. † Boreskov Institute of Catalysis. ‡ Emory University. (1) Sheldon, R. A.; Kochi, J. K. Metal-Catalyzed Oxidations of Organic Compounds; Academic Press: New York, 1981; Chapter 3. (2) Parshall, G. W.; Ittel, S. D. Homogeneous Catalysis: The Applications and Chemistry of Catalysis by Soluble Transition Metal Complexes, 2nd ed.; Wiley: New York, 1992; pp 151-261. (3) Suslick, K. S. In ActiVation and Functionalization of Alkenes; Hill, C. L., Ed.; Wiley: New York, 1989; pp 219-236. (4) Jorgensen, K. A. Chem. ReV. 1989, 89, 431-458. (5) Meunier, B. Chem. ReV. 1992, 92, 1411-1456. (6) Pope, M. T. Heteropoly and Isopoly Oxometalates; Springer-Verlag: Berlin, 1983. (7) Hill, C. L.; Prosser-McCartha, C. M. Coord. Chem. ReV. 1995, 143, 407-455 and references therein.

(8) Hill, C. L.; Khenkin, A. M.; Weeks, M. S.; How, Y. In ACS Symposium Series on Catalytic SelectiVe Oxidation; Ogama, S. T., Hightower, J. W., Eds.; American Chemical Society: Washington, DC, 1993; Chapter 6, pp 67-80. (9) Katsoulis, D. E.; Pope, M. T. J. Am. Chem. Soc. 1984, 106, 27372739. (10) Baker, L. C. W.; Glick, D. C. Chem. ReV. 1998, 98, 3-49. (11) Finke, R. G.; Rapko, B.; Saxton, R. J.; Domaille, P. J. J. Am. Chem. Soc., 1986, 108, 2947-2960. (12) Hill, C. L.; Brown, R. B. J. Am. Chem. Soc. 1986, 108, 536-538. (13) Rong, B. C.; Pope, M. T. J. Am. Chem. Soc. 1992, 114, 2932-2938. (14) Khenkin, A. M.; Hill, C. L. J. Am. Chem. Soc. 1993, 115, 81788186. (15) Kholdeeva, O. A.; Grigoriev, V. A.; Maksimov, G. M.; Zamaraev, K. I. Top. Catal. 1996, 3, 313-325. (16) Zhang, X.; Pope, M. T. J. Mol. Catal. A: Chem. 1996, 114, 201208.

10.1021/ic0000653 CCC: $19.00 © 2000 American Chemical Society Published on Web 08/04/2000

Dimeric Titanium-Containing Polyoxometalate NaClO,8,14 and O2 in the presence of different reductants.27-29 More than 50 papers have addressed the oxygenation or oxidation of a wide variety of substrates in the presence of POM derivatives, and this literature has been recently reviewed.7,8,15,30,31 Since H2O2 is one of the most attractive oxidants from both economic and environmental perspectives,1,2,7,8 Ti(IV)-containing POMs are of particular interest because they can be regarded as homogeneous oxidatively stable models32,33 for TS-1 and Ticontaining mesoporous materials, including Ti-MCM-41, which are highly efficient heterogeneous catalysts for the oxidation of organic substrates by H2O2.34-40 Despite the number of mechanism and structure-activity studies of TS-1 and related materials, the mechanism of their catalytic action and the nature of the active oxidizing species still remain a matter of discussion. Several Ti(IV)-containing POMs are known.41-53 These include (17) Schwegler, M.; Floor, M.; van Bekkum, H. Tetrahedron Lett. 1988, 29, 823-826. (18) Kuznetsova, N. I.; Detusheva, L. G.; Kuznetsova, L. I.; Fedotov, M. A.; Likholobov, V. A. Kinet. Katal. 1992, 33, 516-523. (19) Kholdeeva, O. A.; Maksimov, G. M.; Fedotov, M. A.; Grigoriev, V. A. React. Kinet. Catal. Lett. 1994, 53, 331-338. (20) Neumann, R.; Gara, M. J. Am. Chem. Soc. 1995, 117, 5066-5074. (21) Yamase, T.; Ishikawa, E.; Asai, Y.; Kanai, S. J. Mol. Catal. A: Chem. 1996, 114, 237-245. (22) Ishikawa, E.; Yamase, T. J. Mol. Catal. A: Chem. 1999, 142, 6176. (23) Zhang, X.; Chen, Q.; Duncan, D. C.; Campana, C. F.; Hill, C. L. Inorg. Chem. 1997, 36, 4208-4215. (24) Kholdeeva, O. A.; Maksimovskaya, R. I.; Maksimov, G. M.; Zamaraev, K. I. React. Kinet. Catal. Lett. 1998, 63, 95-102. (25) Faraj, M.; Hill, C. L. J. Chem. Soc., Chem. Commun. 1987, 14871489. (26) Neumann, R.; Abu-Gnim, C. J. Am. Chem. Soc. 1990, 112, 60256031. (27) Neumann, R.; Levin, M. In Dioxygen ActiVation and Homogeneous Catalytic Oxidation; Simandi, L. I., Ed.; Elsevier: Amsterdam, 1991; pp 121-126. (28) Mizuno, N.; Hirose, T.; Tateishi, M.; Iwamoto, M. Chem. Lett. 1993, 1839-1842. (29) Kholdeeva, O. A.; Grigoriev, V. A.; Maksimov, G. M.; Fedotov, M. A.; Golovin, A. V.; Zamaraev, K. I. J. Mol. Catal. A: Chem. 1996, 114, 123-130. (30) Neumann, R. Prog. Inorg. Chem. 1998, 47, 317-370. (31) Okuhara, T.; Mizuno, N.; Misono, M. AdV. Catal. 1996, 41, 113252. (32) Talsi, E. P.; Babushkin, D. E. J. Mol. Catal. A: Chem. 1996, 106, 179-185. (33) Bortolini, O.; Campello, C.; Di Furia, C. F.; Modena, G. J. Mol. Catal. 1982, 14, 63-73. (34) Clerici, M. G. Appl. Catal. 1991, 68, 249-261. (35) Clerici, M. G.; Ingallina, P. J. Catal. 1993, 140, 71-83. (36) Maspero, F.; Romano, U. J. Catal. 1994, 146, 476-482. (37) Bellussi, G.; Cazati, A.; Clerici, M. G.; Maddinelli, G.; Millini, R. J. Catal. 1992, 133, 220-230. (38) Antonelli, D. M.; Ying, J. Y. Curr. Opin. Colloid Interface Sci. 1996, 1, 523-529. (39) Thomas, J. M. J. Mol. Catal. A: Chem. 1997, 115, 371-377. (40) Corma, A. Chem. ReV. 1997, 97, 2373-2419. (41) Tourne´, C. C. R. Acad. Sci., Ser. C 1968, 226, 702-704. (42) Tourne´, C.; Tourne´, G. Bull. Sci. Chim. Fr. 1969, 1124-1136. (43) Ho, R. K. C.; Klemperer, W. G. J. Am. Chem. Soc. 1978, 100, 67726774. (44) Knoth, W. H.; Domaille, P. J.; Roe, D. C. Inorg. Chem. 1983, 22, 198-201. (45) Maksimov, G. M.; Kuznetsova, L. I.; Matveev, K. I.; Maksimovskaya, R. I. Koord. Khim. 1985, 11, 1353-1357. (46) Maksimov, G. M.; Kustova, G. N.; Matveev, K. I.; Lazarenko, T. P. Koord. Khim. 1989, 15, 788-796. (47) Detusheva, L. G.; Fedotov, M. A.; Kuznetsova, L. I.; Vlasov, A. A.; Likholobov, V. A. IzV. Acad. Nauk, Ser. Khim. 1997, 914-920. (48) Chen, Y.; Liu, J. F. Polyhedron 1996, 15, 3433-3436. (49) Domaille, P. J.; Knoth, W. H. Inorg. Chem. 1983, 22, 818-822. (50) Keana, J. F. W.; Ogan, M. D. J. Am. Chem. Soc. 1986, 108, 79517957. (51) Lin, Y.; Weakley, T. J. R.; Rapko, R.; Finke R. G. Inorg. Chem. 1993, 32, 5095-5101. (52) Yamase, T.; Ozeki, T.; Sakamoto, H.; Nishiya, S.; Yamamoto, A. Bull. Chem. Soc. Jpn. 1993, 66, 103-108.

Inorganic Chemistry, Vol. 39, No. 17, 2000 3829

Figure 1. Polyhedral representation of the monomeric Keggin-type POM PTiW11O405- (2), the µ-hydroxo dimer [(PTiW11O39)2OH]7- (H1), and their pH-dependent interconversions. The TiO6 octahedra are shaded.

monosubstituted and disubstituted Keggin-type POMs, including peroxo complexes,41-49 Walls-Dawson-type POMs,50 and more complicated but commonly observed anhydride forms of POMs, such as Si2Ti6W18O7714-,51 Ge2Ti6W18O7714-,52 P4Ti6W32O13236-,54 and P8W60Ti12O24236-.54 The latter compounds are members of an emerging class of POMs that involve condensationdimerization through substituted addenda atoms (both d0 and dn, n > 1).51-55 The tetraalkylammonium salts of PTiW11O405- (2) were found to catalyze alkene oxidation by H2O2 in two-phase systems17,19 but not in homogeneous MeCN.18,19 Recently, we have found that some samples of 2 show high catalytic activity in the oxidation of methyl phenyl sulfide by H2O2.24 We have studied a relationship between the POM countercation and catalytic activity in the above reaction and found that protons are needed to form a peroxo POM that is active in MPS oxidation.56 Here we report on the preparation and characterization of the new titanium(IV)-containing µ-hydroxo dimeric heteropolytungstate [(PTiW11O39)2OH]7- (H1) formed in acidic MeCN from the well-known Keggin POM 2 (Figure 1), the monomerdimer interconversion chemistry, and the role of peroxo Ti POMs in thioether oxidation by H2O2. (53) Crano, N. J.; Chambers, R. C.; Lynch, V. M.; Fox, M. A. J. Mol. Catal. A: Chem. 1996, 114, 65-75. (54) Nomiya, K.; Arai, Y.; Takashi, M.; Takayama, T.; Weiner, H.; Nagata, T.; Widergan, J. A.; Finke, R. G. Inorg. Chim. Acta 2000, 300-302, 285-304. (55) Papers on POMs that involve dimerization through d0-substituted addenda atoms: (a) Finke, R. G.; Droege, M. W. J. Am. Chem. Soc. 1984, 106, 7274-7277. (b) Edlund, D. J.; Saxton, R. J.; Lyon, D. K.; Finke, R. G. Organometallics 1988, 7, 1629-1704. (c) Mu¨ller, A.; Krickemeyer, E.; Dillinger, S.; Meyer, J.; Bo¨gge, H.; Stammler, A. Angew. Chem., Int. Ed. Engl. 1996, 35, 171-173. (d) Day, V. W.; Klemperer, W. G.; Schwartz, C. J. Am. Chem. Soc. 1987, 109, 60306044. (e) Kim, G.-S.; Zeng, H.; Rhule, J. T.; Weinstock, I. A.; Hill, C. L. J. Chem. Soc., Chem. Commun. 1999, 1651-1652. (f) Kim, G.-S.; Zeng, H.; VanDerveer, D.; Hill, C. L. Angew. Chem., Int. Ed. Engl. 1999, 38, 3205-3207. Papers on POMs that involve dimerization through dn-substituted (n > 1) addenda atoms: (g) Zonnevijlle, F.; Tourne´, C. M.; Tourne´, G. F. Inorg. Chem. 1982, 21, 2751-2757. (h) Martin-Frere, J.; Jeanin, Y. Inorg. Chem. 1984, 23, 3394-3398. (i) Randall, W. J.; Weakley, T. J. R.; Finke, R. G. Inorg. Chem. 1993, 32, 1068-1071. (j) Fedotov, M. A.; Detusheva, L. G.; Kuznetsova, L. I.; Likholobov, V. A. Zh. Neorgan. Khim. 1993, 38, 515-525. (k) Wassermann, K.; Palm, R.; Lunk, H.-J.; Fuchs, J.; Steinfeld, N.; Sto¨sser, R. Inorg. Chem. 1995, 34, 5029-5036. (l) Wei, X.; Bachman, R. E.; Pope, M. T. J. Am. Chem. Soc. 1998, 120, 10248-10253. (m) Kortz, U.; Jeanin, Y. P.; Te´ze´, A.; Herve´, G.; Isber, S. Inorg. Chem. 1999, 38, 3670-3675. (56) Kholdeeva, O. A.; Maksimov, G. M.; Maksimovskaya, R. I.; Kovaleva, L. A.; Fedotov, M. A. React. Kinet. Catal. Lett. 1999, 66, 311-318.

3830 Inorganic Chemistry, Vol. 39, No. 17, 2000 Experimental Section Materials. Acetonitrile (Fluka, HPLC quality) was dried and stored over activated 3 Å molecular sieves. Methyl phenyl sulfide (MPS), methyl p-tolyl sulfide (MTS), and the corresponding sulfoxides were purchased from Fluka and used without additional purification. Tetran-butylammonium hydroxide (TBAOH) (1.0 M solution in MeOH, Aldrich) and trifluoromethanesulfonic acid (TfOH, Merck) were commercial samples and were titrated prior to use. Hydrogen peroxide (35%) was concentrated in vacuo to 68-86% and was titrated iodometrically prior to use. All the other compounds were the best available reagent grade and were used without further purification. The Na salt of PW11O397- was prepared by adding 1/11 mol of H3PO4 to 1 mol of H3PW12O40 dissolved in water and adjusting the pH to 4.0 with NaOH (the 31P NMR spectrum in H2O shows a single line at -10.7 ppm57) (stoichiometry: 11H3PW12O40 + H3PO4 + 84NaOH f 12Na7PW11O39 + 60H2O). H5TiPW11O40 was synthesized by the electrodialysis method described previously (31P NMR in H2O: -13.68 ppm).58,59 Instrumentation and Methods. Gas chromatographic analyses were performed using a gas chromatograph equipped with a flame ionization detector and a 15 m × 0.3 mm SE-30 capillary column (Ar, 100-200 °C, 10 °C/min). GC-MS analyses of organic products were conducted using an LKB-2091 instrument. FAB-MS spectra of the POMs were obtained in a 2-nitrophenyl octyl ether (NPOE) matrix using the negative ion mode on a JEOL SX102/SX102A/E mass spectrometer. 31P, 17O, and 183W NMR spectra were recorded at 161.98, 54.24, and 16.67 MHz, respectively, on an MSL-400 Bru¨ker spectrometer. Chemical shifts, δ, were referenced to 85% H3PO4, H2O, and 1 M aqueous Na2WO4, respectively. Chemical shifts upfield from the reference are reported as negative. In 31P NMR measurements, a secondary standard, 0.1 M aqueous H4PVMo11O40, was used. Its δ, -3.70 ( 0.03 ppm relative to concentrated H3PO4, was corrected for the solution magnetic susceptibility difference. The error in measuring δ was in the range of (0.04, (3, and (0.1 ppm for 31P, 17O, and 183W NMR spectra, respectively. Electronic absorption spectra were run on a Shimadzu UV-300 spectrophotometer using 1 cm thermostated quartz cells. Infrared spectra were recorded for 0.5-1.0 wt % samples in KBr pellets or for MeCN solutions in KBr cells on a Specord 75 IR or a Nicolet 510M FTIR spectrometer. All the cyclic voltammetric measurements were performed at 25 °C under argon using a BAS CV-50W voltammetric analyzer, a three-electrode cell, a glassy-carbon working electrode, a platinum auxiliary electrode, and a Ag/AgNO3 (0.01 M in MeCN) reference electrode. TBAPF6 (0.1 M) was used as the supporting electrolyte. The sweep rate in all cases was 100 mV/s. C, H, and N analyses were performed at Atlantic Microlab Inc., Norcross, GA. All other elemental analyses were performed at E&R Microanalytical Laboratory, Inc., Corona, NY. TBA7[(PTiW11O39)2OH] (TBA salt of H1). Synthesis of the TBA salt of H1 was carried out as follows: To 9 mmol of H5PTiW11O40 in 45 mL of water was added 31.5 mmol of TBABr dissolved in 30 mL of H2O. The resulting mixture was vigorously stirred (5-7 min); then the white precipitate was separated by centrifugation, washed with H2O, and dried at 70 °C. The resulting solid was dissolved in MeCN, filtered to remove insoluble impurities, and reprecipitated by the addition of a 5-fold v/v excess of water. The resulting white solid was separated, washed, dried at 70 °C until the weight remained constant, and then placed under vacuum (1-2 Torr) for 4 h at 100 °C. No additional weight was lost (yield: 36 g; 98% based on initial POM). The TBA salt of H1 is reasonably soluble (up to about 0.05 M) in dry MeCN. The number of TBA cations, determined by the ignition of the product salts at 600 °C, was ca. 3.5 per P atom. Anal. Calcd for C112H253N7O79P2Ti2W22: C, 18.78; H, 3.56; N, 1.37; O, 17.64; P, 0.86; Ti, 1.34; W, 56.45. Found: C, 19.11; H, 3.58; N, 1.37; P, 0.66; Ti, 1.16; W, 56.33. IR (1200-400, cm-1): δ 1076, 971, 891, 815, 655, 594, 515. 31P NMR (57) Maksimov, G. M.; Maksimovskaya, R. I. Polyhedron 1996, 15, 42754276. (58) Maksimov, G. M.; Maksimovskaya, R. I.; Kozhevnikov, I. V. Zh. Neorgan. Khim. 1992, 37, 2279-2286. (59) Kulikova, O. M.; Maksimovskaya, R. I.; Kulikov, S. M.; Kozhevnikov, I. V. IzV. Akad. Nauk SSSR, Ser. Khim. 1991, 1726-1732.

Kholdeeva et al. (0.05 M in dry MeCN at 20 °C): δ -12.76 (∆ν1/2 ) 2 Hz). 17O NMR (0.05 M in MeCN at 45 °C): δ 767, 743, 727 (WdO), 547 (W-OTi), 417, 408, 403, 394 (W-O-W), 62 (PO4); the approximate intensity ratio of terminal to bridging oxygen atoms is 1:2. 183W NMR, (0.05 M in MeCN at 20 °C): -δ 89.5, 98.5, 99.2, 101.5, 105.4, 108.7, with the approximate intensity ratio of 2:2:1:2:2:2. The negative-anion FABMS spectrum (m/z 4000-8000) gave m/z 5408 ( 5 as the most intense peak. This is strong evidence of a dimeric structure and may be attributed to [H1 + 6H - 4O]-. The molecular ion peak [TBA7H1 H]- at m/z 7163 ( 4 observed in the FAB-MS spectrum of the TBA salt of H1 is in agreement with the value of 7165 calculated for the molecular weight of the TBA salt of H1. The fragmentation patterns are dominated by loss of multiple TBA (m/z 242), O (m/z 16), and WO3 (m/z 232) units. Potentiometric titration of H1 (0.05 mmol) in MeCN (5 mL) with 1 M methanolic TBAOH, carried out as described in the literature,11,60 shows a sharp breakpoint at 1 equiv of OH-, indicating a [H+]/[H1] ratio of 1.0. Repeated attempts to grow diffraction-quality single crystals of the TBA salt of H1 failed. Many solvents, including alcohols and water, cleave the dimeric structure, and monomeric Keggin POM species crystallize in cubes from mixtures of MeCN with these solvents. TBA5PTiW11O40 (TBA salt of 2). The TBA salt of 2 was prepared by two different procedures. The first procedure is a slight modification of that in the literature.47 Ti(SO4)2 (6 mmol, 1 M solution in 2 M H2SO4) was added to an aqueous solution (20 mL) of Na7PW11O39 (6 mmol). After the pH of the resulting solution was adjusted to 5.6 by adding small aliquots of NaHCO3, the product was precipitated by adding a 5-fold molar excess of TBABr (30 mmol). The white precipitate was separated, purified, and dried as described for H1. The elemental analyses were satisfactory and consistent with the formula TBA4.6Na0.4PTiW11O40. The second procedure provides a route to the pure TBA5 (Na+-free) salt of 2. In this procedure, 3 equiv of methanolic TBAOH was added to a MeCN solution of H1. A white solid, precipitated by the addition of a 5-fold v/v excess of water to the MeCN solution, was separated, washed, dried at 70 °C, and analyzed (satisfactory elemental analysis as TBA5PTiW11O40). The samples from both procedures had identical IR and 31P NMR spectra. IR (1200-400 cm-1): 1089, 1062, 960, 886, 705 (weak), 660 (weak), 593, 513. 31P NMR (dry MeCN): δ -13.34 (∆ν1/2 ) 1 Hz). 17O NMR (0.1 M in MeCN at 45 °C): δ 963 (TidO), 723, 716, 706 (WdO), 536, 523 (W-O-Ti), 420, 407, 389 (W-O-W), 70 (PO4). 183W NMR (0.1 M in MeCN at 20 °C): -δ 62.4, 94.7, 103.7, 108.1, 112.7, 119.6, with the approximate intensity ratio 2:2:1:2:2:2. The negative-anion FABMS spectrum exhibited no peaks in the m/z 4000-8000 range, consistent with the presence of only a monomeric Keggin structure. Potentiometric titration with methanolic TBAOH revealed no acid protons. TBA5PTiW11O41 (II). The peroxo complex II was prepared by the reaction of the TBA salt of 2 with 30% aqueous H2O2 in MeCN at 20 °C ([H2O2]/[2] ) 15) over a period of 3 days. The yellow solid, which had precipitated when H2O was added to the MeCN solution, was centrifuged, collected, washed twice with water, and dried in a vacuum desiccator overnight. 31P NMR (dry MeCN): δ -12.98 (∆ν1/2 ) 18 Hz). The peroxo complex is quite stable (only 10% decomposition to 2 is observed after it is stored in the dark at room temperature for 6 months). The [II]/[2] ratio can be easily quantified by the integration of 31P NMR signals at -12.98 and -13.34 ppm, respectively. Catalytic Oxidation of Thioethers. Oxidations of MPS and MTS were carried out in a thermostated glass reactor equipped with a magnetic stirrer, a reflux condenser, and an inlet for argon. The initial rates of thioether (0.5 M) oxidation by H2O2 (0.1 M, added as 86 or 35 wt % aqueous solution) at 25 °C in MeCN were used to assess the relative catalytic activities of the different Ti-containing POMs (5 × 10-4 M). Aliquots (1 mL) were taken, and the H2O2 consumption vs time was determined by iodometric titration in an AcOH-H2O (70/ 30, v/v) medium. The organic reactants and products were monitored by gas chromatography as a function of time. Products were identified from their 1H NMR and mass (GC-MS) spectra. The yields of sulfoxide and sulfone products as well as the conversion of initial sulfides were quantified by GC using biphenyl as the internal standard. (60) Hou, Y.; Hill, C. L. J. Am. Chem. Soc. 1993, 115, 11823-11830.

Dimeric Titanium-Containing Polyoxometalate

Inorganic Chemistry, Vol. 39, No. 17, 2000 3831

Figure 2. Typical curve for the potentiometric titration of H1 (0.05 mmol) in MeCN (5 mL) with 1.0 M methanolic TBAOH. The potentials are relative to a standard Ag/AgCl electrode. Stoichiometric Oxidations by the Peroxo POMs. To further investigate the mechanism of the catalytic H2O2-based thioether oxidation, we examined stoichiometric reactions of the peroxotitanium complexes. Peroxo complex II does not react with MPS at room temperature over a period of several days (no changes were observed in the 31P NMR and UV-vis spectra of II, and no conversion of MPS was detected). To transform II (0.02 M, δ -12.98) to the active complex I (δ -12.43), we added 1 equiv of acid (HClO4 or TfOH) as a 0.2 M solution in MeCN. Then MPS or MTS (0.1 M) was added to I, and the disappearance of the signal at δ -12.43 and the appearance of the peaks of the monomeric POMs at -13.34 and -13.44 ppm were monitored by 31P NMR. The disappearance of I (maximum at 25 000 cm-1) was simultaneously monitored using UV-vis spectroscopy. The interaction of I with MPS, giving MPSO, was confirmed by GC.

Results and Discussion Synthesis and Hydrolytic Chemistry of [(PTiW11O39)2OH]7(H1). We observed previously that tetralkylammonium salts of PTiW11O405- (2) precipitated from aqueous solutions with different acidities, showing significantly different catalytic activities in MPS oxidation by H2O2 in MeCN.24,56 To understand this phenomenon, we have studied the effect of the POM preparation procedure on the composition of the catalyst and its activity in this typical polar organic medium. The presence of the µ-hydroxo dimer, H1, a new complex, was first indicated by a potentiometric titration of the TBA salt of the titanopolytungstophosphate (precipitated from H5PTiW11O40) in MeCN with methanolic TBAOH. Figure 2 shows a typical potentiometric titration curve for 1, with a sharp breakpoint at 1 equiv of OH-, indicating [H+]/[H1] ) 1.0. Further addition of TBAOH results in a slow response of the glass electrode, consistent with cleavage of the µ-oxo dimer, a result borne out by all the additional experimental data (vide infra). In general, the dimer (H1) forms upon acidification of 2 by strong acids (e.g., HClO4 or triflic acid). Several lines of evidence convincingly establish that the µ-hydroxo dimeric formulation of H1 is correct, that H1 and 2 are interconvertable by reversible, pH-dependent reactions (addition of base forms 2; addition of acid forms H1), and that the dimer can be isolated in both basic (Ti-O-Ti ) 1) and protonated (Ti-OH-Ti ) H1) forms. Each line of experimental evidence will now be succinctly presented. The reader should refer to the Experimental Section for details. First, the infrared spectra show that the Keggin unit in 2 is retained in H1 (Figure 3 compares the two spectra in the informative 2200-400 cm-1 region). The solid (KBr pellet) and solution (MeCN) spectra of H1 are practically identical,

Figure 3. IR spectra of (a) H1 and (b) 2. Both samples are at 0.5-1.0 wt % in KBr.

indicating that the same structure is present in both phases. The spectra for H1 and 2 differ significantly in two locations. The triply degenerate PO4 stretching mode is split in 2 (1089 and 1062 cm-1) but not in H1 (1076 cm-1), indicating a higher symmetry around the central PO4 tetrahedron in H1 than in 2. This phenomenon, first elaborated by Rocchiccioli-Deltcheff and Thouvenot61 and probed in several subsequent studies,43,44,47,62-66 gives a useful indication of the distortion induced by the replacement of a (M(VI)dO)4+ (M ) W,Mo) unit in the parent Keggin structure with other atoms or groups, such as (Ti(IV)d O)2+. The band at 665 cm-1 in H1 is most likely the asymmetric stretching fundamental of the Ti-O-Ti unit in this dimer for two reasons: first, this band is absent in 2, and second, this band is in the same region as the M-O-M stretches in β-(TBA)7.5H6.5Si2Ti6W18O77 (Ti-O-Ti at 700 cm-1),51 K19H9P4Ti6W32O132 (Ti-O-Ti at 683 cm-1),53 and TBA12H4P4W30Nb6O123 (Nb-O-Nb at 683 cm-1),55a,b a complex recently confirmed by X-ray crystallography to be a tri-µ-oxo dimer55e Second, the number, δ range, and intensity ratios of the 17O and 183W NMR resonances (see Experimental Section) confirm the retention of the Keggin structure of the PTiW11O40 units in H1. The 17O signal for the terminal oxygen atom on Ti is not seen in the spectrum of H1 (Figure 4a) but appears (δ 963) after the addition of 3 equiv of TBAOH in MeOH (Figure 4b). Previously, the signal of the terminal TidO in 2 was observed in aqueous solution at 919 ppm.47 In addition, the relative line widths in the 17O NMR spectra of H1 and 2 support the dimeric formulation of H1. Third, the negative-ion FAB-MS spectrum recorded in the range of m/z 4000-8000 in an inert 2-nitrophenyl octyl ether (61) Rocchiccioli-Deltcheff, C.; Thouvenot, R. J. Chem. Res. Synth. 1977, 2, 46-47. (62) Combs-Walker, L. A.; Hill, C. L. Inorg. Chem. 1991, 30, 4016-4026. (63) Kim, G.-S.; Hagen, K. S.; Hill, C. L. Inorg. Chem. 1992, 31, 53165324. (64) Khenkin, A. L.; Hill, C. L. J. Am. Chem. Soc. 1993, 115, 81788186. (65) Kim, G.-S.; Judd, D. A.; Hill, C. L.; Schinazi, R. F. J. Med. Chem. 1994, 37, 816-820. (66) Zhang, X.; Chen, Q.; Duncan, D. C.; Campana, C.; Hill, C. L. Inorg. Chem. 1997, 36, 4208-4215.

3832 Inorganic Chemistry, Vol. 39, No. 17, 2000

Figure 4. 17O NMR spectra of H1 (0.05 M in MeCN) (a) before and (b) after addition of 3 equiv of methanolic TBAOH.

Figure 5. Negative ion FAB-MS spectrum of H1 dissolved in a NPOE matrix.

(NPOE) matrix (Figure 5) strongly supports the presence of a dimer in H1. The most intense peak at m/z 5408 ( 5 may be attributed to [H1 + 6H - 4O]-. The observed molecular ion peak [TBA7H1 - H]- at m/z 7163 ( 5 is in excellent agreement with the molecular weight of 7165 calculated for TBA7[(PTiW11O39)2OH] (TBA7H1). This peak is completely absent in the negative-anion FAB-MS spectrum of 2, consistent with the presence of only a monomeric Keggin unit. The fragmentation patterns for both 1 and 2 are dominated by the loss of multiple TBA (m/z 242), O (m/z 16), and WO3 (m/z 232) units characteristic of the FAB mass spectra of POMs.67,68 The spectra for each complex were recorded at least three times and were of excellent reproducibility, establishing that no monomerdimer conversions were taking place during desorption or other aspects of the analytical process.69 Fourth, 31P NMR spectra, under a range of conditions, afford corroboration of the implications of the infrared, 17O NMR, 183W NMR, and FAB-MS data and provide additional information (67) Finke, R. G.; Droege, M. W.; Cook, J. C.; Suslick, K. S. J. Am. Chem. Soc. 1984, 106, 5750-5751. (68) Mizuno, N.; Hirose, T.; Tateishi, M.; Iwamoto, M. J. Mol. Catal. 1994, 88, L125-L131. (69) It is interesting to note that H5PTiW11O40, the synthetic precursor to 1, appeared to be monomeric in aqueous solution according to FABMS data. However, it appears to be dimeric in the solid state, since the intense band at 665 cm-1 is observed in its KBr pellet IR spectrum. Previously, this band was erroneously assigned to the Ti-O-W bonds.58

Kholdeeva et al. about the titanopolytungstophosphates in MeCN. A peak at -13.34 ppm was previously assigned to the TBA salt of the monomer, 2.24 The µ-hydroxo dimer, H1, shows a single peak at -12.76 ppm in dry MeCN. This signal is broadened upon dilution (∆ν1/2 ) 2, 4, 11, and 25 Hz at 0.05, 0.005, 0.0025, and 0.00125 M, respectively). Such broadening in dry MeCN may be explained by the slow proton mobility on the polyanion surface in polar nonaqueous solvents.11,70-72 Dilution of H1 to beyond ca. 0.001 M results in the reversible generation of 2 (a peak at -13.44 ( 0.04 ppm). 31P NMR facilitated the direct observation of µ-oxo (hydroxo) dimer formation and cleavage. Stepwise addition of small quantities of a strong acid (HClO4 or TfOH) gives rise to 2 in a different protonation state and additional species, each with distinct resonances (Figure 6). Initially, the addition of acid causes the appearance of a new NMR signal at -13.44 ppm at the expense of the signal at -13.34 ppm (Figure 6b). Note that the chemical shifts of both resonances are not effected by the addition of small quantities of H2O. Upon the further addition of acid, the intensity of the signal at -13.44 ppm decreases, and a new signal appears at about -13.2 ppm, which then moves downfield (Figure 6c). When 1.5 equiv of acid is added, the chemical shift coincides with that of H1 (Figure 6d,e). These results are entirely consistent with the initial protonation of the TidO oxygen (δ -13.34) in forming P(TiOH)W11O404- (H2, δ -13.44), since the protonation of Ti-O-W bridges should be expected to result in a downfield shift of the 31P resonance, analogous to the shift in vanadium-containing POMs.73 Protonation of the TidO bond prior to Ti-O-W formation was proposed previously for the Ti-substituted Well-Dawson POMs50 as well as for H5PTiW11O40.58 It is not possible to isolate pure H2, since, on the basis of 31P NMR spectra, this species reacts further. The formation of Ti-O-Ti units upon the acidification of hydrolyzed Ti(IV) compounds is well known and used in sol-gel processing.74,75 Furthermore, the µ-oxo dimer A-β-(TBA)7.5H6.5Si2W18Ti6O77 forms in acidic medium (pH 1-4).51 By analogy, it is reasonable to suggest that the monomeric species H2 is a key intermediate in the transformation of 2 to the dimer, 1. Protonation of the Ti-O-Ti bridge occurs upon the further addition of acid, which results in the downfield shift of the 31P resonance to -12.76 ppm at 1.5 equiv of H+ per 2. The protonated dimer, H1, then predominates. The plausible mechanism for the formation of H1 upon acidification of 2 may be described by eqs 1-3

P(TiO)W11O395- + H+ h P(TiOH)W11O394-

(1)

P(TiOH)W11O394- + P(TiOH)W11O394- h [(PTiW11O39)2O]8- + H2O (2) [(PTiW11O39)2O]8- + H+ h [(PTiW11O39)2OH]7- (3) Further addition of acid causes a downfield shift of the signal of H1 to -12.18 ppm, evidently due to the protonation of the (70) Harmalker, S. P.; Pope, M. T. J. Inorg. Biochem. 1986, 28, 85-95. (71) Klemperer, W. G.; Shum, W. J. Am. Chem. Soc. 1977, 99, 35443545. (72) Klemperer, W. G.; Shum, W. J. Am. Chem. Soc. 1978, 100, 48914893. (73) Kholdeeva, O. A.; Maksimovskaya, R. I. Zh. Neorg. Khim. 1992, 37, 1349-1354 and references cited therein. (74) Doeuff, S.; Dromzee, Y.; Taulelle, F.; Sanchez, C. Inorg. Chem. 1989, 28, 4439-4445. (75) Schneider, M.; Baiker, A. Catal. Today 1997, 35, 339-365.

Dimeric Titanium-Containing Polyoxometalate

Inorganic Chemistry, Vol. 39, No. 17, 2000 3833

Figure 7. Cyclic voltammetric behavior of 2 (3 mM in MeCN) (a) before acid is added and after addition of (b) 0.5 and (c) 1.5 equiv of TfOH. Potentials are reported relative to Ag/AgNO3 (0.01 M in MeCN). All samples contained 0.1 M TBAPF6 as the supporting electrolyte.

Figure 6. 31P NMR spectra of 2 (5 mM) (a) before acid is added and after the addition of (b) 0.75, (c) 1.25, and (d) 1.5 equiv of HClO4. (e) 31P NMR spectrum of H1 (5 mM). All spectra were run in MeCN at 20 °C.

Ti-O-W oxygen atoms. The line widths indicate that proton exchange between the different dimeric species still remains fast on the 31P NMR time scale. Equation 2 predicts that the addition of H2O to the system would shift the monomer-dimer equilibrium in favor of the monomer, H2. 31P NMR experiments confirm that this is precisely what happens. The addition of water to 2 does not effect the chemical shift and width of its resonance, whereas the 31P resonance of H1 shifts upfield and narrows with increasing H2O concentration, and the signal at -13.44 ppm for the protonated monomer, H2, appears. This signal also appears upon dilution of pure H1. Thus, the process opposite to that observed upon the acidification of 2 occurs. Water competes with the µ-oxo dimer for the protons. Since the rate of proton exchange between protonated and unprotonated species increases with H2O concentration, signal narrowing is observed. H2 becomes the predominant form at [H1]0 ) 0.0025

M in 9:1 v/v MeCN:H2O. Also consistent with eqs 1-3 is the observation that upon the addition of 3 equiv of methanolic TBAOH to H1, 2 is formed immediately in very high selectivity. Note that the tri-µ-oxo dimer TBA12H4P4W30Nb6O123 is readily cleaved by 6 equiv of OH- to yield TBA9P2W15Nb3O62.55b,c In contrast to 1, 2 is quite stable against bases in MeCN solution; the addition of 3 equiv of TBAOH causes no change in its 31P NMR spectrum after several days at 25 °C. The fifth and final line of experimental observations on the nature of the titanopolytungstophosphate system concerns its electrochemical behavior. The cyclic voltammograms of H1 and 2, examined under conditions as close as possible to those in the catalytic thioether oxidations (vide infra), confirm the pHdependent interconversions of these two POMs. First, upon the addition of 0.5 equiv of triflic acid to 2, which should result in the formation of H2, parallel changes in two reduction peaks were observed. While the peak at E1/2 ) -2.12 V decreases by a factor of 4.9, the peak at E1/2 ) -1.23 V increases in a parallel mode by a factor of 4.8 (Figure 7, traces a and b). The changes in the cyclic voltammograms are consistent with the assignment of the peaks at -2.12 and -1.23 V as the one-electron reduction/ oxidation of 2 and H2, respectively. The two individual peaks for the reduction of 2 and H2 (instead of one peak shifting upon addition of acid) are due to slow proton exchange between 2 and H2 on the CV time scale, which is also consistent with the observation of two individual 31P NMR signals assigned to 2 and H2 as well (Figure 6b). The further addition of H+ (up to 1.5 equiv) results in the appearance of two new peaks at E1/2 ) -0.90 and -0.71 V (Figure 7, trace c) presumably corresponding to the one-electron reduction of 1, H1, and possibly other protonated derivatives of 1 that are formed at higher pH. Unlike 2 and H2, dimeric derivatives (1, H1, etc.) were shown from 31P NMR to be in a fast proton exchange in MeCN solution. As a result, more specific assignment of the latter reduction peaks becomes difficult. The addition of 0.5 equiv of OH- (as methanolic TBAOH) results in the reversible decrease in the peaks at E1/2 ) -0.90 and -0.71 V. The complete disappearance of these peaks is observed after the addition of 1 equiv of OH-. Under identical conditions, the cyclic voltammetry experiment with the TBA salt of H1 prepared independently (and not in situ by acidification of 2) also gave the peaks at E1/2 ) -0.90 and -0.70 V. Again, these disappeared upon the

3834 Inorganic Chemistry, Vol. 39, No. 17, 2000 Table 1. Thioether Oxidation by H2O2 in the Presence of H1 and 2a catalyst

substrateb

initial ratec (×106 M s-1)

H1 H1 2 H1 2 H1 H1 H1

MTS MTS MTS -e -e MPS MPS MPSO

2.9 ( 0.2 5.0 ( 0.3d 0.45 ( 0.02 0.0063 0.0016 3.0 ( 0.3 4.5 ( 0.3d 1.7 ( 0.1

Reaction conditions: 2.5 × 10-4 M H1 or 5 × 10-4 M 2, 0.5 M substrate, 0.1 M (added as 86 wt % solution) H2O2, in MeCN, at 25 °C. b MTS, MPS, and MPSO are methyl p-tolyl sulfide, methyl phenyl sulfide, and methyl phenyl sulfoxides, respectively. c Loss of H2O2 with time ) (-[H2O2]/dt)o. d With 0.1 M (added as 35 wt % solution) H2O2. e In the absence of the substrate. a

addition of 1 equiv of OH- (relative to the number of moles of the titanium atoms), which supports the assignments of these peaks as the one-electron reduction of 1 and H1. Data from these five experimental techniques plus potentiometric titration all lead to the same mutually supportive conclusions that 1 in MeCN solution is a previously unknown dimer with a Ti-O-Ti bridge that forms from the dehydration condensation of two Keggin monomers (H2) and that acid and base clearly convert the monomers to the µ-oxo dimers and visa versa. Furthermore, 1 is stable in its protonated form (H1) with a Ti-OH-Ti bridge, and a protonated form of the monomer, H2, is an observable intermediate in these interconversions leading to eqs 1-3 as a summary of this hydrolytic chemistry. Precipitation of the TBA salts of 2 in the pH range of 0.8-2.0 yields various mixtures of H1, 2, and H2. Catalytic Oxidation of Thioethers by H2O2 in the Presence of H1. Active and Inactive Peroxo Complexes. As we reported previously,24 thioethers can be efficiently oxidized by H2O2 in the presence of titanopolytungstophosphates at room temperature to produce the corresponding oxygenated products (sulfoxides, RR′SO, and sulfones, RR′SO2) in practically quantitative yield based on both H2O2 and substrate consumed. In contrast, the selectivity for sulfoxide, while very high, is not quantitative as in other systems76-78 because sulfoxide is oxidized further to sulfone at approximately 2% the rate of the thioether oxidation. The selectivity for sulfone starting with 100% sulfoxide is effectively quantitative. In the absence of any titanopolytungstophosphate, neither the thioether MPS nor the sulfoxide MPSO is oxidized by H2O2 (even 86%) at ambient temperature. The fact that this system (titanopolytungstophosphates/MeCN/H2O2) results in no alkene oxidation even at elevated temperature (5060 °C) is corroborating evidence that the Keggin units of the titanopolytungstophosphates are stable under these conditions. This follows the patterns of PW11O397- and other polyoxotungstates quickly converted by H2O2 to polyperoxotungstates, which result in facile alkene epoxidation by H2O2 because these peroxo complexes that form in situ are all very good epoxidation catalysts.17,18,79,86-88 Table 1 summarizes the reactivity data for oxidation of two representative thioethers, MPS and MTS, by H2O2 in MeCN at room temperature in the presence of H1 and 2. It was determined that neither light nor the presence of O2 affects the rates of these catalytic oxidation reactions. The addition of 20 mol % (relative (76) Gall, R. D.; Faraj, M.; Hill, C. L. Inorg. Chem. 1994, 33, 50155021. (77) Gall, R. D.; Hill, C. L.; Walker, J. E. J. Catal. 1996, 159, 473-478. (78) Gall, R. D.; Hill, C. L.; Walker, J. E. Chem. Mater. 1996, 8, 25232527.

Kholdeeva et al. to H1) of the radical scavenger hydroquinone has no demonstrable affect on the rate or other characteristics of these reactions. These facts collectively indicate that the radical chain component of these reactions is small, if it exists at all, and/or that the propagation steps in the chain are significantly faster than the quenching reactions (hydroquinone + freely diffusing radical intermediates). Table 1 indicates that H1 and 2 differ significantly in both their catalytic oxidation and their peroxide disproportionation activity, with H1 being more reactive than 2 for both reactions. The rate of H2O2 decomposition in the absence of thioether is about 2-3 orders of magnitude lower than the rate of H2O2 consumption in the presence of thioether. At 20 °C, [H1] ) 5 × 10-3 M and [H2O2] ) 0.1 M; 44% of H2O2 is decomposed in 8 days. The catalytic activity of the TBA titanopolytungstophosphates, H1 and 2, for thioether oxidation by H2O2 also correlates with the rate of each transformation into the active peroxo complexes (vide infra). A systematic study of H1 or 2 in the absence of the substrate (i.e., the full catalytic system without the substrate) reveals key features of these oxidations and the presence of two peroxotitanium complexes. Several lines of evidence establish the nature of these two peroxotitanium complexes and, to some extent, their role in substrate oxidation. First, treatment of an 0.005 M MeCN solution of 2 with a 30-fold molar excess of H2O2 results in a slow (about 24 h) formation of an orange peroxo complex (II) (31P NMR singlet at -12.98 ppm), whereas treatment of H1 under the same conditions results in the more rapid formation (about few hours) of a second peroxo complex (I; 31P NMR singlet at -12.43 ppm). Second, both I and II have very similar UV-vis spectra, with maxima at λ ) 400 and 395 nm, respectively, which in turn are very similar to the spectrum for the peroxo complex P(TiO2)W11O395- known in aqueous solution (392 nm,  ) 1600 M-1 cm-1),46,47 consistent with peroxo groups in all three compounds. Third, both FAB-MS and infrared spectra show the disappearance of the dimeric structure on treatment of H1 with H2O2. Dinuclear titanium µ-oxo/µ-peroxo complexes are precedented,32,80,81 and the cleavage of µ-oxo linkages in dimeric POMs by H2O2 is also well precedented.55a Fourth, as in the case of the reactant POMs H1 and 2, the peroxo complexes I and II can be converted into each other by the addition of 1 equiv of H+ or OH- (Figure 8). The stepwise addition of H+ to II results in the downfield shift and narrowing of the 31P resonance (δ (∆ν1/2, Hz) -12.88 (12), 12.77 (16), 12.66 (10), and 12.43 (4) at 0.2, 0.4, 0.6, and 1.0 equiv of H+, respectively). This confirms that I and II differ in their proton content and are in the fast exchange on the NMR time scale. The higher acidity and consequently higher electrophilicity of I versus II is consistent with their relative reactivities toward the nucleophilic thioether substrates. Fifth, I is more reactive and less stable than II. Thioether oxidation proceeds rapidly when 31P NMR reveals the presence of I but very slowly when 31P NMR reveals the presence of only II. This is in accordance with our preliminary data.24 It is well known that acid additives (Brønsted and Lewis) enhance the catalytic activity of peroxotitanium porphyrin complexes82 and titanium silicalites.34-37 Clearly, protons are needed to generate the active forms of the peroxo species in these systems. The ambiguity here is that acids of all kinds also directly activate (79) Duncan, D. C.; Chambers, R. C.; Hecht, E.; Hill, C. L. J. Am. Chem. Soc. 1995, 117, 681-691. (80) Muhlebach, J.; Muller, K.; Schwarzenbach, G. Inorg. Chem. 1970, 9, 2381-2390. (81) Schwarzenbach, D. Inorg. Chem. 1970, 9, 2391-2397. (82) Ledon, H. J.; Varescon, F. Inorg. Chem. 1984, 23, 2735-2737.

Dimeric Titanium-Containing Polyoxometalate

Inorganic Chemistry, Vol. 39, No. 17, 2000 3835

Figure 9. 31P NMR spectra of 2 containing 1 equiv of H+ (a mixture of 2 and H2) (a) before and (b) 10 min after addition of H2O2. The -13.34 and -13.44 ppm peaks in (a) are assigned to 2 and H2, respectively, and the -12.43 and -13.34 ppm peaks in (b) are assigned to the peroxo complex I and 2, respectively.

Figure 8. 31P NMR spectra of 2 (0.005 M) 36 h after the addition of (a) 20 µL of 77% H2O2 with no added H+, (b) 0.5 equiv of H+, and (c) 1.0 equiv of H+, showing the formation of some η2-peroxotitanium complex (II). 31P NMR spectra of H1 (0.005 M) 4.5 h after addition of (d) 20 µL of 77% H2O2 and (e) 1 equiv of OH- (TBAOH in MeOH).

H2O2 for oxidations of nucleophilic substrates, including thioethers.83,84 The argument in favor of thioether oxidation via a peroxotitanium complex is that the rate of the thioether oxidation increases with the increase of water concentration (Table 1), whereas the opposite effect would be expected if the reaction were catalyzed by H+. In the latter case, water and thioether would compete for protons, and the concentration of the RR′SH+ species (mechanism via nucleophilic oxygen transfer) would decrease. Sixth, the rate of the peroxo complex formation correlates with the catalytic reactivity of different Ti POMs; both the rate of peroxo formation and substrate oxidation are significantly higher for H1 than for 2. This is not surprising, given that the TidO group in Ti(IV) porphyrins has long been known to be inert toward numerous reagents.85 Importantly, the rate of the peroxotitanium complex formation is higher for the protonated monomer (H2) than for H1. Specifically, when H2O2 is added to solutions containing mixtures of either H2 and 1 or H2 and 2, H2 (δ -13.44) readily transforms into the active peroxo complex, I (δ -12.43), whereas the peaks for 2 (δ -13.34) or (83) Madesclaire, M. Tetrahedron 1986, 42, 5459-5495. (84) Banard, D.; Bateman, L.; Cunneen, J. I. In Organic Sulfur Compounds; Kharasch, N., Ed.; Pergamon Press: Oxford, 1961; pp 229-247. (85) Fournary, P.; Guilard, R.; Fontesse, M. J. Organomet. Chem. 1976, 110, 205-217.

1 plus H1 (δ -12.8 to -13.2) remain unchanged (Figure 9). These relative rates of peroxotitanium formation (H2 > 1 > 2) suggest that all peroxotitanium species form via H2. This is also consistent with eq 2 and the observation that the rate of catalytic thioether oxidation increases with increasing water concentration. Stability of the Titanopolytungstophosphate Catalysts. As indicated above, many Keggin POMs rapidly degrade in the presence of aqueous H2O2. If a second chlorocarbon phase containing a phase transfer catalyst (PTC), usually a quaternary ammonium cation, is also present, then the polyperoxometalates formed in situ are extracted into the organic layer where they function as effective catalysts for epoxidation with H2O2.79,86,87 In contrast, the Keggin titanopolytungstophosphates in MeCN with H2O2, the focus of this study, do not degrade under same conditions of time and temperature. Interestingly, while this titanopolytungstophosphate/MeCN/H2O2 system does not epoxidize alkenes, these same POMs in H2O/chlorocarbon twophase systems do.17,19 To understand this difference, we examined H1 by 31P NMR in both homogeneous MeCN and biphasic CH2ClCH2Cl/H2O in the presence of a 100-fold molar excess of H2O2 over H1. One can see from Figure 10a that the only signal observable in homogeneous MeCN is that of the active peroxo complex I (δ -12.43); no products derived from degradation of the Keggin POM unit are detectable in the range of +7 to -22 ppm even after storing the solution at 100 °C for 24 h or increasing the peroxide molar excess to 500 ([H2O2]/[H1] ) 500). Significantly, the Keggin unit is largely destroyed in aqueous solution under identical conditions (Figure 10b). Under the conditions for alkene epoxidation in the two-phase system (60 °C, [H2O2]/ [catalyst] ) 500, PTC ) tetrahexylammonium chloride),17,19 the Keggin POM is degraded. H2O clearly results in the breakdown of titanium-containing POMs, like many other (86) Aubry, C.; Chottard, G.; Platzer, N.; Bre´geault, J.-M.; Thouvenot, R.; Chauveau, F.; Huet, C.; Ledon, H. Inorg. Chem. 1991, 30, 44044415. (87) Salles, L.; Aubry, C.; Thouvenot, R.; Robert, F.; Doremieux-Morin, C.; Chottard, G.; Ledon, H.; Jeannin, Y.; Bre´geault, J.-M. Inorg. Chem. 1994, 33, 871-878.

3836 Inorganic Chemistry, Vol. 39, No. 17, 2000

Kholdeeva et al.

Figure 10. 31P NMR spectrum of (a) H1 (0.01 M) in MeCN and (b) H5PW11TiO40 in H2O (pH 1.8), both 24 h after the addition of a 100fold molar excess of H2O2.

POMs, in two-phase catalytic H2O2-based oxidation systems and, thus, facilitates generation of the lower nuclearity species, which are the actual catalysts. Another group has reported that the alkene oxidation by H2O2 occurs in MeCN in the presence of 2.21,22 In light of these results and the 31P NMR data reported earlier,22 however, it is quite likely that the observed activity might be due to impurities in the catalyst and/or species derived from the degradation of 2. With the above data in hand, we now turn to the final fate of the catalyst. The active peroxotitanium species, I, persists from its initial high-yield generation from H1 until all the H2O2 has been consumed. At this point, its characteristic yellow color and 31P NMR signal (δ -12.43) disappear, and H2 or a mixture of H2, 1, and H1 is formed. The 31P NMR for H1 shifts slightly in the latter stages of reaction because deprotonation of H1 by the increasing sulfoxide product and the H2O byproduct (derived from the consumed H2O2) is building up. Both sulfoxide and H2O are known to deprotonate the highly acidic unsubstituted heteropoly acid POMs (i.e., those with only Mo(VI) or W(VI) addenda atoms). The situation is less clear (pKa values not yet adequately established) for heteropoly acids substituted with one or more lower valent metal ions, including Ti(IV) or Nb(V). In one case, the actual sulfoxonium (protonated sulfoxide) salts of the POMs were isolated and shown to be stable under very similar reaction conditions (peroxide oxidations of thioethers catalyzed by vanadopolymolybdophosphates in MeCN).76 In a control experiment, we added some product sulfoxide (MPSO) to the substrate-free system (H1/H2O2/MeCN), and the same shifts as those in the spent catalytic reactions were seen (upfield shifts of the 31P NMR resonance of H1). In confirmation of this deprotonation effect, the same shifts in the 31P NMR peaks were also seen when small quantities of H2O or MeOH were added to the substrate-free system. The 31P NMR signals remain relatively sharp throughout in all these reactions because the proton transfer (POM protonation/deprotonation) reactions are considerably faster than the NMR time scale (i.e., the fast exchange limit is operational).

Figure 11. 31P NMR spectra of II (a) before the addition of H+, (b) after the addition of 1 equiv of H+, and after the addition of MPS recorded after (c) 2, (d) 15, (e) 25, and (f) 60 min. All spectra were run in MeCN at 20 °C. [II] ) 0.02 M, and [MPS] ) 0.1 M.

Stoichiometric Oxidation of Thioether Substrate by the Reactive Peroxotitanium Complex. The best evidence for a particular step (elementary process) in a mechanism derives from

Dimeric Titanium-Containing Polyoxometalate finding experimental conditions where that step only is experimentally operable. Under such conditions, the step can be directly observed and studied. In this catalytic thioether oxidation chemistry (thioether/H1/H2O2/MeCN), we sought final confirmation that the peroxotitanium complex, I, was the active oxidant under turnover conditions. Thus, we sought to generate it cleanly and examine its direct interaction (oxidation) with one molecule of thioether substrate, in this case, MPS. While I is not sufficiently stable for satisfactory isolation and independent study, it can be generated from II in very high yield by protonation. Again, II is generated directly by reaction with H2O2 and 2, and it is unreactive to thioether at ambient temperature in MeCN solution. The TBA salt of II has actually been known since 1985,46 and its diisopropylammonium salt was characterized by X-ray crystallography, which confirms it contains a “side-on” or η2-peroxometal group.89 η2-Peroxometal groups, like that in II, are typically minimally reactive and electrophilic;21,82,90,91 they exhibit weak O-O stretching bands in the IR (no dipole moment change is possible for such symmetrically bound ligands), and they are activated by protons and Lewis acids, as is II, which converts to I. Figure 11 shows through 31P NMR the generation of I from the addition of 1 equiv of H+ to II and the disappearance of I upon the addition of the thioether MPS. The direct reaction between I and MPS was also monitored by UV-vis spectroscopy. Note that I slowly decomposes in MeCN; however, the rate of its decomposition estimated from UV-vis and NMR data is about 20 times slower than the rate of its consumption in the presence of MPS. The splitting of the 31P NMR signal of the parent Ti-POM into two lines is most likely due to its protonation upon the addition of the acid (see Figure 6). The disappearance of I directly parallels the formation of sulfoxide from the thioether, as detected by GC. Since I is the only active oxo transfer species in the reaction, it is reasonable to infer that it is reacting directly with the thioether substrate to generate the sulfoxide product. To the best of our knowledge, this is the first demonstration of a direct stoichiometric reaction between a peroxotitanium complex and an organic substrate. At this stage of research, the exact structure of the active peroxo complex, I, and the site of the activating proton remain uncertain. The proton may be on a peroxo oxygen atom (η1hydroperoxo complexes A and B, respectively, or η2-hydroperoxo complex C), or it may be on the Ti-O-W bridge (η2peroxo complex D) (Figure 12). The fact that the chemical shift of I is not affected by added H2O argues against structure D because bases should result in the upfield shift of its 31P resonance, as in the case of H1. Hydroperoxotitanium species types A, B, and C have been proposed as the active intermediates in the commercially useful H2O2-based oxidations catalyzed by TS-1 and related systems;34-37,92-95 however, compelling spectroscopic evidence in favor of one of these structures has (88) Dengel, A. C.; Griffith, W. P.; Parkin, B. C. J. Chem. Soc., Dalton Trans. 1993, 2683-2688. (89) Yamase, T.; Qzeki, T.; Motomura, S. Bull. Chem. Soc. Jpn. 1992, 65, 1453-1459. (90) Mimoun, H.; Postel, M.; Casabianca, F.; Fisher, J.; Mitschler, A. Inorg. Chem. 1982, 21, 1303-1306. (91) Dickman, M. H.; Pope, M. T. Chem. ReV. 1994, 94, 569-584.

Inorganic Chemistry, Vol. 39, No. 17, 2000 3837

Figure 12. Possible structures for the active peroxotitanium complex.

yet to be obtained. In the present study the active hydroperoxotitanium complex was detected spectroscopically for the first time. Further spectroscopic and kinetic studies are in progress to clarify the structure of this active species and the precise mechanism of its interaction with the thioether substrates. Conclusions Three main conclusions can be drawn from the present work. First, four forms of the monosubstituted Ti(IV)-containing Keggin POMs have been generated and investigated in MeCN solution, namely, the µ-hydroxo dimer [(PTiW11O39)2OH]7(H1) and its conjugate base [(PTiW11O39)2O]8- (1) and two monomers PTiW11O405- (2) and its conjugate acid P(TiOH)W11O394- (H2). The relative concentrations of these four species depend on the concentrations of both H+ and H2O. Acetonitrilesoluble TBA salts of H1 and 2 can be easily isolated in pure form. Second, the catalytic activity of these titanium-containing POMs in thioether oxidation by H2O2 correlates with their activity in H2O2 decomposition and with the rates at which they form the peroxotitanium complexes. Rates for peroxotitanium formation and thioether oxidation are H2 > H1 > 2. Third, protons are needed to form the active peroxo species, which is most probably a hydroperoxotitanium complex, P(TiOOH)W11O395-. This complex reacts directly with thioether substrates under both stoichiometric and catalytic conditions. Acknowledgment. The authors thank K. F. Obzherina (Boreskov) for IR measurements and a referee, Richard G. Finke, for his suggestions. The authors also appreciate the assistance of I. V. Khavrutskii (Boreskov). This work was supported in part by Russian Foundation for Basic Research (Grant 96-03-34215) (O.A.K., G.M.M., R.I.M., and L.A.K.) and in part by the U.S. National Science Foundation and the U.S. Army Research Office (C.L.H.). IC0000653 (92) Hulea, V.; Moreau, P.; Di Renzo, F. J. Mol. Catal. 1996, 111, 325332. (93) Geobaldo, F.; Bordiga, S.; Zecchina, A.; Giamello, E.; Leofanti, G.; Petrini, G. Catal. Lett. 1992, 16, 109-115. (94) Karlsen, E.; Schoffel, K. Catal. Today 1996, 32, 107-114. (95) Zecchina, A.; Bordiga, S.; Lamberti, C.; Ricchiardi, G.; Scarano, D.; Petrini, G.; Leofanti, G.; Mantegazza, M. A. Catal. Today 1996, 32, 97-106.